Re: Load Testing

From: Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com>
To: Ashish Karalkar <ashishka(at)synechron(dot)com>
Cc: Abdul Rahman <abr_ora(at)yahoo(dot)com>, pgsql-general(at)postgresql(dot)org
Subject: Re: Load Testing
Date: 2009-02-13 18:58:49
Message-ID: dcc563d10902131058v787c23bfv5fc7d12d7f21ed1b@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Fri, Feb 13, 2009 at 4:48 AM, Ashish Karalkar <ashishka(at)synechron(dot)com> wrote:
> Ashish Karalkar wrote:
>>
>> Abdul Rahman wrote:
>>>
>>> Hi all,
>>>
>>> Can any body tell me about tool for PostgreSQL load testing preferably
>>> freeware.
>>>
>>> Regards,
>>> Abdul Rehman.
>>>
>> I am not sure its a freeware or not but looks promising
>>
>> http://bristlecone.continuent.org/HomePage
>>
>>
>> --Ashish
>>
> And ofcourse the PGbench which is freeware:
>
> http://www.postgresql.org/docs/current/static/pgbench.html

Yep. pgbench is kind of my basic acceptance testing benchmark. If
you've got a 16 core 128G ram machine hitched onto a 100+15k5 SAS disk
san array and you're getting 20 tps on pgbench there's not much use in
running other benchmarks until you figure out what's so wrong.

It's also good for applying burn in loads over long periods. Nothing
like a week of running pgbench to find problems with RAID controllers,
drives, memory, cpus, cooling, power supplies or kernels. I had a
kernel bug on a server last year that took about 12 hours of heavy
pgbench to show up. Had a bad RAID controller that took 24 to 36
hours of pgbench to hang.

Plus, pgbench has the ability to run custom SQL for benchmarking, so
it's an easy way to build a custom test.

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Ray Stell 2009-02-13 19:59:31 8.3 doc issue
Previous Message Raymond O'Donnell 2009-02-13 18:31:54 Re: Remote Connection